Results of the First Archaeological Excavation and Environmental Survey at Fengtai, Qinghai Province, PR China, in 2001
Xu Xinguo, Xining, Wang Wei, Peking, Mayke Wagner, Karl-Uwe Heussner, Pavel E. Tarasov, Bettina Griess, Berlin, Niu Shishan, Song Jiangning, Peking, Xiao Yongming and Cai Linhai, Xining
Schlagwörter: NW-China/Bronzezeit/Kayue-Kultur/Xindian-Kultur/Siedlung Keywords: NW China/Bronze Age/Kayue Culture!Xindian Culture/Settlement llpeo11,ie111Hb1. e cJ1 oea: C3 Kumaü/3nox a 6po1-13bi/KyJ1 bmypa KG/03/K y llbmypa CUHbOfl.Hb!lloceJ1eHue
Objectives
The Depa1tment of Eurasian Archaeology of the German Archaeological Institute deals with human occupation and land use history in the steppe belt of Eurasia (e.g. Iran, Black Sea Coasts, Kazakhstan and Tuva). Qinghai Province, NW China, situated at the south eastern margin of the Great Eurasian Steppe, has been chosen as one of the key areas for joint studies in order to evaluate the role of anthropogenic and clirnatic factors in envi 1 ronmental changes • Qinghai Province is located at the northeastem edge of the Qinghai-Tibetan Plateau between ca 32° and 39° N and ca 90° and 103° E. This is a region with complex topogra phy and hydrology. Mountain ridges and high plateaus elevating from about 2,500 to 4,500 m above sea level dominate the region. The present day clirnate is transitional from sub-humid to semi-arid with mean annual precipitation fluctuating between 250 and 600 mm per year and a strong water deficit. The main part of precipitation comes during May-September and is associated with the Pacific Monsoon. Up to 40% of the annual 2 precipitation has been known to fall in a single storm • This together with a spread of loess sediment and intensive land use has led to strong erosion activity in the region. According to up-to-date knowledge the neolithic habitation of Qinghai started at about 4000 BC in the eastem pait of the Province along the Y ellow River and its tributar ies. The features of the ancient cultures are very distinct and different from those ob served in Central and Eastern China. However, at certain periods the cultural similarities to the adjacent territories, the modern Provinces Gansu and Sichuan, the Autonomous Regions Xinjiang and Tibet and even Southern Siberia, Kazakhstan and Iran are striking. The driving forces behind the rise and fall of these cultures, and human mobility and mi-
1 The excavation campaign in June 2001 was organised by the Archaeological Institute of Qinghai Province (team members: Xu Xinguo, Xiao Yongming, Cai Linhai, Liu Xiaohe and Zhang Changshou) and the Archaeological Institute of the CASS (Wang Wei, Niu Shishan and Song Jiangning) with participation of the Department of Eurasian Archaeology DAl (M. Wagner, B. G1iess, K.-U. Heussner). Zhao Zhijun (Archaeological Institute of the CASS) collected plant macrofossil remains, D. Jaekel (Berlin) investigated geomorphology of the area and Jin Guiyun (Jinan/Berlin) collected samples for pollen analysis from cultural layers and natural sequences. The field campaign continued in March 2002, when M. Ullrich and K.-U. Heussner (Berlin) made a cartographic plan of the site (Scale 1: 500). We thank the cultural administrations of Qinghai Province and Huzhu County for their generous support of the work. 2 Derbyshire et al. 1991. 86 Xu Xinguo u. a. gration phenomena in the region are still poorly understood for lack of sufficient archaeo logical information. The term "Kayue Culture" was coined in 1958 as an idiom for the Bronze Age in the Qinghai region in order to distinguish it from the Bronze Age in the neighbouring Gansu 3 Province, where it was named "Siwa Culture" • The eponymous site Kayue (Ch'ia Yao) 4 was discovered by the Swedish geologist Johan Gunnar Andersson in 1923 . Since that time approximately 1,700 Bronze Age sites attributed to the Kayue Culture have been found in central and eastem Qinghai Province. Their dating to the latter half of the second and the first half of the first millennium BC5 is mainly based on ceramic vessel typolo gies. The Kayue Culture is supposed to have coexisted with the largely contemporaneous Nuomuhong Culture to the west in the Qaidam Basin and with the Xindian Culture to the east downstream the Y ellow River in modern Gansu Province. More specific points rela ted to the Nuomuhong Culture still remain largely within the realm of speculation för 6 only the Nuomuhong site itself was partly excavated in 1959 • The early settlement histo ry of the Qaidam Basin before the dramatic desertification therefore stays one of the most challenging topics of future investigation which remains cannot be touched upon here. 7 The Xindian Culture, in particular its ceramic production, is much better known • Only very recently have the cultural features of the Kayue Culture together with the few avail able radiocarbon dates been discussed to some extent and put into the context of the entire 8 Bronze Age development of Northwest China by Shui T. . Despite the large number of discovered sites our knowledge about the culture is limited. Same basic information was 9 obtained by the partial excavation of seven graveyards • So far no Kayue settlement has been investigated systematically. Tue historical and palaeoenvironmental records from China suggest both lang- and short-term climate changes which are regarded as the main factor having influenced set tlement dynamics since the Mid-Holocene. Basically the economy of the region is very much dependent on climate. However, human impact on landscape dynamics at different scale and degree can be traced through the whole history of human habitation of this area. Today the politically initiated and govemment-promoted intense economic development of Qinghai Province dominates all aspects of land use. The results of archaeological and palaeoenvironmental studies conducted in the Fengtai Valley, which is characteristic of a distinct type of landscape in Qinghai Prov ince, i. e. mountain steppe, are important for an understanding of the history of settlement and landscape dynamics in the area. They also provide the opportunity to build up retro spective models of landscape developments back to the Neolithic. At the same time, this knowledge, together with the planned investigation and documentation of recent pro cesses (e. g. mass relocation, population dynamics) in the area, provides a basis for the modeling of future scenarios of landscape development and land use strategies. In this
3 Xu 1988, 35. 4 Andersson 1925, 18; 22f. 5 Mei J. J. quotes current opinions about Kayue dates which all dwell on the sarne sparse material; Mei 2000, 64 Ann. 25. 6 Zhongguo wenwu dituji, Qinghai fence 1996, 96. 7 Recent summaries in Pak 1996, 287 - 299 and Shui 2001, 116- 146. 8 Shui 2001. 9 Wagner 2001. Survey at Fengtai 2001 87
R U S S
17
.. . .. ;,I"'·< . 45°N
' PACIF'.IC
. 30°N 0 OCEAN
··~; · 15°N
Fig. 1. Position of Fengtai site.
paper we present the main results of the füst joint research project conducted at Fengtai site, Qinghai Province, in 2001.
Site setting
The Bronze Age settlement near Fengtai village, 46 km north of the provincial capital Xining (Fig. l) was discovered by members of the Archaeological Institute of Qinghai Province in the course of systematic archaeological surveys in the early 1980s. lt was declared a cultural heiitage zone in 1986. On the basis of red grit tempered potsherds collected from the surface, the site was attributed to the Kayue Culture. According to the Atlas of Chinese Cultural Relics (Qinghai volume) the estimated size of the site is about 2 10 4,200 m and the thickness of the cultural layer is 0.5 m . In July 2000 the Archaeological Institute of Qinghai Province guided an archaeo logical field excursion for colleagues from the Archaeological Institute of the Chinese Academy of Social Sciences (CASS, Peking) and the Department of Eurasian Archae ology of the Ge1man Archaeological Institute (DAI, Berlin). The sites visited were either already excavated and known from publications, sites currently under excavation, or sites registered and protected but not yet surveyed. In Fengtai, which belongs to the last group, the joint research team collected a large number of various artefacts attributed to the Kayue Culture. Moreover, due to road construction along the south-eastem margin of the
IO Zhongguo wenwu dituji, Qinghai fence 1996, 89. 88 Xu Xinguo u. a.
Fig. 2. Fengtai, view on the terrace complexes. site seetions were opened, exposing to view layered eultural deposits of about three me ters in height. The stratigraphie observations and riehness of stray finds lead to the as sumption that the Fengtai site is a weil preserved long-term dwelling plaee providing suffieient arehaeologieal substanee to merit close examination. Our first and most essential task was to establish a ehronologieal framework. By pottery analogues alone we eould not narrow down the age of the settlement to a eenten nial seale. In order to understand the entire time range of this dwelling site, it was first of all neeessary to get a sequenee of absolute datings from a seetion representing all eultural layers from the youngest to the oldest. The exeavation eampaign was therefore primarily devoted to stratigraphie studies. Seeondly, we sought to gain insight into house eonstrue tion teehniques and materials. Detailed information about the Kayue vernaeular arehitee ture would not only be a substantial addition to Bronze Age studies in Northwest China but would equally provide the neeessary parameter for the preparation of subsequent ex tensive exeavations.
Results of the exeavation
Profile stratigraphy
Spread over an area of about 54,000 m2 potsherds were found on the surfaee, in varying density eovering the whole alluvial fan whieh opens south-eastwards to the main valley of the Beiehuan River (Fig . 2). The alluvial fan eonsists of three terraee eomplexes (A, B and C). Aeeording to oral information obtained from peasants the modern terraeing of the C
E F G H
1.0 -
E ..c 1• c age, cal. yr BC 15. 2.0 - Q) 0 ■ One Sigma Range (probability 68.2%)
□ Two Sigma Range (probability 95.4%)
1400 1200 1000 800 3.0 · ~ -.-.·.-.·.-.·. I: ·:w.·r::::· ::: ·x·•:: ~ •7i:",°:; }'·....__. •,-~:";\, ; 1 : : : : it KIA 16590 „ Bln -5414
~ · KIA 165B4
1C3 Surface seil 2[IJJ Modem irrigation channel \~ a , : KI A16588 3~ Mud brick KI A16586 4 c::::::J Yellow sandy clay 11 1 , · .' :=:==::=~~:c:::.:: !ia a ! ! KI A16589 s c::::::J Charcoal band ~ Bin 541 6 ...... : b 6 1,. __ ,. ,· 1 Grey hard clay : ~ i KI A16585 7 OIIl]] Ashy seil 14 ◊ Fired clay pieces smm Ash and clay 15 0 Stones 9 1:-: :-: :- 1 Mixed humified deposits 16 ~ Charcoal particles or layers 10 ~ Ash pit 17 ~ Pottery 11~ Laminated house floors 18 I><] Bones 12 1·.::,·.--:-·.:_J Alluvial sand 19 ,. ,~ Urne particles 13 fZZZJ Base sediment 20 o--- Sampling for 14C dating
Fig. 3. Fengtai, trench TG 1, East Profile. 00 '° 90 Xu Xinguo u. a. whole alluvial fan was done very recently, during the 1980s and the early 1990s. Only modest changes of the previous surface occurred in the upper part of Complex A and in the lowest southem part of Complex B. Therefore we decided to open the main trench TG 1 at the southem rirn of Complex B. Complex A lies at the highest elevation. lt is named "Temple Hili" because of the active Lamaist temple located there (Fig. 2). A gen tle slope leads down from Complex A to Complex C; a natural depression separates ter race Complex A from terrace Complex B and the latter one from Complex C. Several years ago this topographic depression was further deepened by the villagers for irrigation purposes. Due to this soil relocation two sections (south-eastem side of the Temple Hili and north-westem side of Complex B) were opened, allowing us to observe cultural lay ers near the bottom of the channel at a depth of ca. 4.5 m below the sediment surface. For a better understanding of the succession of the cultural layers the profile was cleaned and extended as trench TG 2.
Trench TG 1: East and West Profiles
The main trench TG 1 has a width of 2 m and is approxirnately 8 m long. lt has a general north to south orientation. In this chapter we discuss two profiles, representing strati graphical features of the eastem (Fig. 3) and westem (Fig. 4) banks of trench TGl, re spectively. Despite the fact that the distance between the two profiles is only 2 m, the sediment stratigraphy presented in Figs. 3 and 4 is not similar in every detail. The differ ences are caused by the complex slope geomorphology, sediment dynarnics and the struc ture of the ancient housing. In the upper 1.3 meters of the Bast Profile (Fig. 3, column A) the traces of two shal low irrigation grooves can be seen. These grooves are filled with sediments which do not contain any material suitable for dating. However, based on our experience and the in formation from the villagers they can be explained by the agricultural activities of the last century. Tue uppermost of the ancient cultural layers was found at a depth of about 1 m (Fig. 3, columns A.B). lt is represented by grey hard clay overlying a collapsed mud brick wall. In this "mud brick layer" very destroyed vessel fragrnents and one alrnost complete small jar with a double handle were found. Charcoal particles picked up from the outer wall of the jar were radiocarbon dated to 930- 820 cal BC11 (KIA 16591; Fig. 4, column A). We assurne that this date is the prelirninary age of the mud brick wall (Fig. 5). Tue foundation of this collapsed wall (Fig. 3, column A) is 0.3-0.5 m thick and grounds on rnixed hurnified deposits at about 1.6 m below surface. In order to find out the function of the wall the excavation in the upper part of the slope was extended to 5 m x 5 m. This discontinuous character of the wall does not support the idea that it rnight have been a fortification enclosing the entire settlement. Most likely it was the wall of a single build ing or yard. Tue arnount of ash particles in two ashy layers underlying the wall (Fig. 3, column B - D) decreases downwards. These ashy soil and clay layers together with a yellow sandy clay band probably represent the open surface outside the wall synchronous to it. Below
11 All radiocarbon ages in the text indicated as 'cal BC' are expressed in calibrated years BC. The age range covers two sigma probability (95.4%). For additional information see Fig. 8. 14 ( age, cal. yr BC
G F E D C B A 1400 1200 1000 800
,., , ,- l . Läi 1 ~ ,,..,,
2.0 ' ····························'···········-/ E ..c ö. ■ One Sigma Range (probability 68.2%) •. .. ~ □ Two Sigma Range (probability 95.4%) -- C=:J Surface soil 14 ◊ Fired clay pieces 2 C=:J Modern irrigation channel 15 Cl Stones 3~ Mud brick 16 ~ Charcoal particles or layers 4 C=:J Yellow sandy clay 17 b Pottery 5 C=:J Charcoal band 18 l> Fig. 4. Fengtai, trench TG 1, West Profile...... '-0 92 Xu Xinguo u. a. Fig. 5. Fengtai, trench TG 1, mud brick wall fragments ca 1.5 m below surface. this fonner surface layer we observed several layers of mixed humified deposits disturbed by pits. Tue sediments contain large amounts of burned clay particles, potsherds, animal bones and stones. Fragments of horizontally oriented yellow sandy clay layers can be found at a depth of about 2.2-2.3 m and about 3.1 m (Fig. 3). Similar features can be traced in the West Profile (Fig. 4). We assume that these fragmentary layers represent former dwelling surfaces of different settlement periods. However, their precise dating is a matter of further investigation. Laminated hause floors can be detected in the East Profile (Fig. 3, columns E.F) between 2.7 and 2.9 m andin the West Profile (Fig. 4, columns C-E) at the same depth. High concentration of lime particles within and undemeath the hause floor layer is evi dence of the custom of whitewashing floors. At this level we removed several surfaces of the extremely fine hause floors at a depth of 2.63 to 2.83 m recovering inserted circular hearths and post holes (Fig. 6). The charcoal particles taken from the upper and lower hearths (Fig. 4, column D) were radiocarbon dated to 980-820 cal BC (KIA 16593) and to 1130- 930 cal BC (KIA 16592), respectively. Charcoal was also collected at 2.70- 2.83 m Swvey at Fengtai 2001 93 Fig. 6. Fengtai , trench TG 1, house floor in ca 2.8 m depth with heatth. depth (Fig. 4, column C) and dated to 1260-1000 cal BC (KIA 16594). In the East Profile material for dating was found in the fragmentary preserved house floor layers situated at 3.17 m (Fig.3, column C) and 3.34-3.44 m depth (Fig. 3, column F). The charcoal from these samples was dated to 833-793 cal BC (KIA 16590) and 910-800 cal BC (Bln 5414). The next clearly definable layer is a massive charcoal layer with abundant fragments of fired clay, visible in the East Profile from Column E to Hin a depth of 3.8 to 4.3 m and appearing in the West Profile as a thin band in Columns E and F between 3 .4 and 3. 7 m. One piece of carbonised coniferous wood was dated to 1320-1110 cal BC (KIA 16584). All following buried layers down to the natural loess slope down steeply. They contain a remarkable abundance of ashes and burned wood fragments , preserved in pieces with diameters up to 10 cm and more than 40 year rings and lumps of mud-plaster with imprints of beams and twigs. The pottery is sooty or badly charred but evidently less fragmented than in the upper cultural layers. The lower and deepest cultural deposits do not yield any traces of rarnmed earth house floors , lime-plaster or mud brick rubble. Two 94 Xu Xinguo u. a. more charcoal layers can be regarded as marker horizons in the East Profile: from colurnn B to H between 3.2 and 5.6 m and in colurnn C between 4.0 and 4.4 m depth. Five selec ted charcoal samples have provided the following absolute ages: 1190- 930 cal BC (KIA 16588), 1190- 940 cal BC (KIA 16586), 1190- 970 cal BC (KIA 16589), 1260- 1040 cal BC (KIA 16585) and 1190- 920 cal BC (Bln 5416). Tue last two dates are obtained from samples of mixed pieces of charcoal collected from the same layer in close proxirnity. They were processed in different ways: with conventional measurement in the laboratory of the German Archaeological Institute in Berlin and with AMS technology in Kiel. Tue Berlin lab date corresponds perfectly to the other datings of this group obtained by AMS in Kiel (KIA 16588, 16586, 16589). At the two sigma probability range of 95.4% these dates have an overlap of 150 years. Nevertheless, the date KIA 16585 is close to KIA 16584 which overlaps with the duster of four corresponding datings for 80 years. We assume that these two older dates were processed from old wood particles within the whole assemblage of burned wood, and therefore do not represent the age of this particu lar cultural layer (old-wood effect). Direct comparison of the conventional and AMS radiocarbon dating methods shows the risk of using only AMS dates acquired from bulk material due to a possible age diversity of the samples. Tue undisturbed loess was reached in the upper parts of the terraced trench in colurnn A to C and in the lowermost parts in colurnn G and H. Tue bottom of the deepest cultural layer was reached in colurnn G at 7.12 m beneath the surface. Due to this unex pectedly complex stratigraphy with lots of architectural features the scheduled excavation time was not sufficient to reach natural soil for the full length of the trench. Trench TG 2: South Profile Already during our first survey in 2000 we observed cultural layers in the recently opened sections in the modern irrigation channel between terrace complexes A and B. Ashy soil, potsherds and animal bones were visible in a thick layer 0.5 m above the channel floor approximately 4.6 m below the eastem terrace surface on top of which a high density of ceramic rubble was recorded. Tue brick-red and thick-walled pottery picked out from the section matched the ware which had been collected on the surface. Therefore, a more or less continuous sequence of cultural deposits - comparable to trench TG 1 - was to be expected. Surprisingly, after the section walls had been cleared of a grassy overgrowth, the profile was found to reveal no traces of cultural layers or even stray finds but instead a layer of yellow sandy clay (Fig. 7, layer 2) followed by clean loess with only some bands of light grey palaeo-soil in the lower parts (Fig. 7, layer 1 ). Instead of upward the cultural deposits continued downward. Layer 3 (Fig. 7) with a certain amount of grey clay, fired clay and charcoal particles overlaid a thick horizon with big lumps of fired clay and bur ned wood. Among this conglomerate of badly burned material were big stones and animal bones. Of particular interest was the presence of quite astonishingly weil preserved ceramic vessels (Fig. 11 ). Taking all aspects into account the assemblage of TG 2 indica tes a dwelling place which was destroyed by fire. Since a relocation scenario can be ex cluded we have to consider the question of why the middle zone of the settlement is cov ered by four meters of loess whereas the sedirnent cover at the westem and eastem rim ranges only between 0.3 and 1.5 m and consists of heavily reworked soil. To put it another way, to understand the entire settlement plan the reconstruction of the contempo rary surface as weil as the subsequent processes of site formation is necessary. Tue dig in Survey at Fengtai 2001 95 E L ä. Q) 0 1'1 1 1 1' :, II 1 1 '11 III ~ Loess ', 1 1 III •' •• II 1•1 1 1 1 ,• •• 1 2 c::=J Yellow sandy clay 1 •• ,: •1 1• ,, II II ,'1 •1 1' 3 Grey clay with fired clay and charcoal 1 ~ '1 II 1' 1 •1 II 4 c::=J Cultural horizon ,', II 1 1, 1 '•' 1 II ,, 1 II 1 ••1 5 1·.-::-:.-::-:.-::-:.-:J Alluvial sand 1 II '1' 1 •• 1 3.0 '1 II :1 1 1 1 Humified layers II ,777 II 1 ,1 6 II 1 1 ,1 ,1 1 1 :1: II Charcoal particles or layers 1 1 7 -- 8 • Pieces of hard clay 9 -0 Fired clay 10 ,CJ Stones 11 @ Ceramic pot with perforated bottom 12 ~ Bones 4.0 13 0- Sampling for 14 C dating 2 14 C age, cal. yr BC > < ' , ,: ,: ·= ' o '·~ 3 ■ One Sigma Range (probability 68.2%) .. ...,. :• { ~ ·· O :- ...... □ Two Sigma Range (probability 95.4%) 900 700 500 5.0 rii KIA16587 4 0 ,CJ Bln5413 Fig. 7. Fengtai, trench TG 2, South Profile. '-0 Laboratory No . Material Depth in the Radiocarbon Calibrated age, yr BC (probability Cal ibrated age, yr BC Period °'- sequence, cm age, uncal. yr 68 .2%) (probability 95.4%) BP charcoal (conifers, Pinus Bin 5413* fragments, 5 to 11 rings) 560 2564±36 810-760, 690-660, 620-590, 580-560 820-750, 710-540 Late KIA16587* charcoal (deciduous tree twig) 480-540 2615±25 812-792 830-765 Laie KIA16590 charcoal 317 2645±25 824-800 833-793 Late charcoal (mixture of wood fragments: Populus (?) and Bin 5414 conifers, max 25 rings ) 334-344 2690±34 900-875, 860-850, 845-800 910-800 Laie KIA16591 charcoal 143 2730±25 900-830 930-820 Late KIA16593 charcoal 283 2755±30 920-890, 880-835 980-820 Late KIA16592 charcoal 292 2865±25 1110-1100, 1080-970, 960-940 1130-930 Late 1210-1200, 1190-1170, 1160-1140, KIA16594 charcoal 270-283 2915±30 1130-1040 1260-1240, 1220-1000 Late/AT0 KIA16588 charcoal 496-501 2870±25 1130-970 1190-1180, 1130-930 Early 1190-1170, 1160-1140, KIA16586 charcoal (deciduous tree twig) 506-508 2880±25 1130-1000 1130-970, 960-940 Early 1190-1170, 1160-1140, KIA16589 charcoal 543-557 2885±25 1130-1000 1130-970 Early charcoal (mixture of wood fragments: conifers and Populus- Bin 5416 Sa/ix, max 11 rings ) 546-568 2868±32 1130-970, 960-940 1190-1180, 1130-920 Early KIA16585 charcoal 546-568 2950±25 1260-1230, 1220-1120 1260-1040 Early charcoal (coniferous tree 9 rings, KIA16584 outer 12art) 410-430 2975±25 1260-1120 1320-1110 Early charcoal (mixture of wood fragments: deciduous and Bin 5415 conifers, max 10 rings ) 404-424 2729±36 900-830 970-950, 930-800 Early/ATY Fig. 8. Fengtai, radiocarbon dates. Dates were calibrated by OxCal v 3.5 (Bronk Ramsey, 2000). Atmospheric data from Stuiver et al. (1998). Dates indicated with a star are from TG 2, others from TG 1. The values of d13C are in the normal range for organic samples. ATO - age too old; ATY - age too young. Survey at Fengtai 2001 97 TG 2 could not be continued down to natural sediments because of the existing irrigation channel. Two charcoal samples from TG 2 were dated independently in the laboratories of Kiel and Berlin. Both samples have a clearly distinguishable small time variation (Fig. 8). For AMS we chose a deciduous tree twig and for the conventional method a carbonised Pinus fragment of five to eleven rings. Tue AMS date provides an age of 830-765 cal BC (KIA 16587) matching the range of datings for the upper layers in trench TG 1. The conventional date of 820-540 cal BC (Bln 5413) also falls into this time bracket. The lang range is caused by the plateau of the calibration curve. Therefore we conclude that the assemblage of trench TG 2 corresponds to the younger data cluster of trench TG 1. The litho- and chronostratigraphy of both trenches being compared allow for the following summary: The lower cultural layers in trench TG 1 between 3 and 7 m are cha racterised by a steep slope, a tremendous amount of ash, bumed wood and pieces of mud plaster. The absolute age determination of samples from this group of layers, ranges be tween 1190 to 920 cal BC, i.e. the Early Period, excluding the dates which are regarded as too old due to the old-wood effect (Fig . 8). This sedimentary and chronological cluster has not been found in trench TG 2, perhaps for lack of sufficient depth. In trench TG 1 at a height of over 3 m the profiles show horizontal hause floors with multi-layered mud- and lime-plasters (Fig. 3, columns E.F; 4, columns C-E; about 2.7 to 2.9 m deep). By extending the trench at this level beyond the limits of the hause floor to the surrounding footstep-hardened dwelling surface a grove-like wall-base could be de tected, but no post holes or any other traces of wall construction material were found. By contrast, very distinct remnants of a wall made of adobe were laid open in the first pre historic cultural layer at about 1.5 m depth. The one date directly related to the wall fits the cluster of dates obtained from the horizontal hause floors . Combined with the two dates from TG 2, the dates for the upper set of layers range from 980 to 750 cal BC, giv ing an age orientation for the Late Period (Fig. 8). lt may be, however, that there is a mo re discrete chronological gap between the lime-floors and the adobe wall which does not show clearly in the present data set but can be proved in future campaigns. The periodisation inh·oduced here is not valid for the entire Fengtai site without fur ther proof by large-scale excavation. We use it to sort out and describe the stratigraphic evidence extracted from the two trenches opened in 2001 and as an expression of the fortunate convergence of sedimentary and chronology records. Cerarnic finds Dominating among the objects unearthed from all layers are different kinds of fragmented vessels of coarse tempered reddish-brown ware and fine red ware. The highest concentra tion of potsherds was found in the charcoal layers of trench TG 1, in particular in columns G and Hat a depth of between 5.5 and 6.0 m. Therefore the bigger part of the recorded potsherds yield traces of buming, often to the degree of being completely carbonised. The entire spectrum of reconstructed vessel types in Fengtai is relatively poor. lt is mainly limited to big, thick-walled, bulbous jars with double vertical handles and wide bowls with a wall thickness up to two centimetres. The colour ranges from dark brown to brick red or light brown; light grey or polished dark black ware is extremely rare. A red slip around the mouth, on the outside surface and inside the neck, as weil as on the handle was obviously highly in favour. Many pieces bear a rather carelessly applied paste-like blood- 98 Xu Xinguo u. a. 3 5 6 ~ 2cm Fig. 9. Fengtai, pottery, Early Period (1190 - 920 cal BC), trench TG 1. red slip. Types of plastic decoration that were observed include applications in the shape of buckles, flanges and ridges with fingerprints, furthermore impressions of cords, mats, irregular notches and cuts. This kind of rough ceramics was recorded in all layers from top to bottom without obvious typological variations related to chronology. Some pottery characteristics could nevertheless be distinguished for the early and late periods as mark ed by sediment stratigraphy and absolute chronology as discussed above. At a depth of 5.5 to 5.7 min column Gin the lower position of the largest charcoal layer the complete orifice part of a jar came to light (Fig. 9,8). Its long double handles arch down from a flaring mouth over a curved-in neck to gently looping shoulders. This very peculiar neck curvature combined with long loop-handles finds parallels in vessel 12 13 types from Dahuazhongzhuang, Panjialiang, Banzhuwa Il and Huangzhai 1 • In both assemblages this shape is regarded as significant for an early phase of the Kayue devel opment which is believed to be closely linked with the preceding Qijia Culture. To some 12 Wagner 2001, 48; 52; 41 Fig. 5,4. 13 Shui 2001 , 241. Survey at Fengtai 2001 99 4 a-~ 6 7 JB, ~------F, JQ)-U 11 0- L--....J 2 cm Fig. 10. Fengtai, pottery, Late Period (980 -750 ca! BC), t.rench TG 1. 100 Xu Xinguo u. a. 2 3 5 6 L____J 2cm 0 Fig. 11. Fengtai, pottery, Late Period (980-750 cal BC), trench TG 2. S urvey at Fengtai 2001 101 ■■■■■■■■■■ Fig. 12. Fengtai, pottery, survey trench beside the temple in 0. 3 m depth. 102 Xu Xinguo u. a. degree the stratigraphic position of this jar in Fengtai supports this idea. The most fre quently recorded orifice form, however, is the one with a short, straight neck, slightly bending out towards the mouth (Fig. 9,5 column D at 3.4- 4.4 m; Fig. 9,6.7 column G at 5.4 - 5.7 m). Single or double short handles are applied to the lip and the shoulder. Close to the lip-handle join a ridge with fingerprints is added which only runs along the handle width but not around the whole circumference. This form of rirn-handle with applied fin gerprint-ridge was rare in the upper layers where only few pieces expose a comparable shape but with barely visible remnants of the ridge (Fig. 10,10 column A at 1.9 to 2.1 m). fu the upper layers the more characteristic rirn-handle shape is the straight form with two buckles instead of a ridge (Fig. 10,12 column A at 1.3 m). Another quite representative type of appendage is the ring handle attached to the belly at the point of its maxirnum diameter (Fig. 10,13 column A at 1.3 m). The type of shallow cup with a pronounced concave collar only appeared in the upper layers and is known from Banzhuwa II14 (Fig. 10,7.9). Tue ceramic vessels from trench TG 2, believed to belong to the same chronological period as the upper layers in trench TG, 1 present a complete picture of the jar type (Fig. 11,3.5.6) . Additionally, a unique form of a little cup with a huge flaring mouth that merges into wide handles which seem to be oversized compared to the small volume in a spherical belly bears a dark-red slip and polish inside the neck (Fig.11,4). Painted omaments in black on brick-red ground were met in both trenches and all layers (Fig. 9,1-4; 10,1 - 5). They consist mainly of hatched triangles (Fig. 9 ,2.4; 10,4.5) and meander (Fig. 9,1; 10,3) correlated for instance to Ahatla finds 15 16 and Banzhuwa pottery . A very specific decoration feature for the late period is black or red painting on a white coating (Fig. 10,1.2.6-8; 11,1). m this, markers of the Xindian Culture can be recognised. Tue Xindian influence became even more evident when a group of weil preserved baggy-legged tripods were excavated in a trial section close be side the temple on the highest terrace (Fig. 12). They correlate to the last stage of the Xindian development, the so-called Zhangjiazui Phase which is dated to 1100- 800 BC by Shui T.17. Results of the environmental survey The understanding of the geomorphic processes as weil as the present and past climate and vegetation dynarnics in the Fengtai area are important for the discussion of the ar chaeological finds. The Fengtai V alley is an intermountain depression deepened by water erosion. At present it is being drained by a creek, flowing eastwards to the Beichuan Ri ver. Tue valley is oriented from west to east and has a length of about 3 km and a maxi mum width of about 1.4 km. Thus, the catchment area of 4 km2 is rather small. Within the catchment area the maximum elevation of 2,841.5 m above sea level is registered in the south-westem part of the valley and in the lowermost part near its mouth the altitude is about 2,530 m. Tue Fengtai Valley has a canyon-like floor and very steep slopes, suggest ing intensive erosion processes during recent times. 14 Wagner 2001 , 41 Fig. 5,3.10. 15 Shui 2001 , 231 Fig. 24, M36: l.M53 : l.M51 : 1. 16 Wagner 2001, 40 Fig. 4,9. 17 Shui 2001 , 122-123 Fig. 2; 131. Survey at Fengtai 2001 103 The loess sediments accumulated in the valley cover the underlying Tertiary red clay. The surface of the clay horizon, becoming lubricous after moisture penetration, is highly vulnerable to the loess movement and associated landslides. The earthquakes, which are frequent in the region, might be another factor causing the landslides18. The southern slope of the valley is under a strong influence of linear erosion. lt is cut by seven large gullies. The distance between neighbouring gullies varies from 200 to 300 m. The steeper northem slope has been considerably changed by landslide activity and later erosion processes. Remnants of several big landslides can be observed in the middle as weil as in the lower part of the valley. Two older slides and a younger one with still very steep edges are clearly visible from the excavation place. The of January - T!) , from 14.5 to 16.0 °C (mean temperature of July - Tv11 ), from 940 to 1,300 °C-day (annual sum of mean daily temperatures above 5 °C called sum of growing degree days - GDD5) and from 0.52 to 0.6 (ratio of actual to potential evapotranspiration called moisture index - a). The last two characteristics are important bioclimatic vaiiab les, used in the vegetation modelling to define the climatic boundaries between the prin cipal vegetation types or biomes21. For example, a = 0.65 marks the boundary between cold deciduous (e.g. birch, poplar, larch), cool temperate (e.g. elm) or eurythermic coni ferous (e.g. pine, juniper) arboreal taxa which require a > 0.65 and cool steppe vegetation with a < 0.65. Sirnilarly, Tv11 = 23 °C separates a cool steppe from a warm steppe bi ome22. The definition of the p1incipal biomes in the BIOMEl model suggests that the present-day natural vegetation in Fengtai region should be a cool steppe. However, the modern vegetation of the area reflects a strong anthropogenic influence. Agiicultural and pasture lands dominate the modern landscape. Most of the natural or man-made flat sur- 18 Derbyshire et al. 1991. 19 Leemans/Cramer 1991. 2°Cramer , pers. communication. 2 1 Prentice et al. 1992. 22 Prentice et al. 1992. 104 Xu Xinguo u. a. faces are used for agricultural purposes. Among the plants grown by the villagers are raps, potatoes, wheat, cabbage, cannabis, linen, and other herbaceous plants. Wheat fields appear up to at least 2,800 m. Land, which is not suitable for agriculture like steep moun tain slopes is used as pastures for numerous sheep, goats, cows, yaks, donkeys and horses. The plant communities of the pastures are represented by various steppe and meadow plants, including grasses, sedges, Artemisia and other Asteraceae species, as weil as Api aceae, Fabaceae, Rosaceae and Polygonaceae. During the past decades central and local governments have paid special attention to the forestation of the area. Small patches and isolated stands of poplar trees grow in the Beichuan River V alley and on mountain slopes exposed to the north. There we found few young spruce (Picea) trees of 0.5 to 1.0 m height. Trees have been planted in the local environments with better moisture conditions, protected from direct sunshine and strong wind. Among the other arboreal taxa growing in the vicinity of Fengtai, elm (Ulmus pumila) and sea buckthom (Hippophae rhamnoi des) should be mentioned. Forest patches of birch (Betula platyphylla), spruce (Picea crassifolia) and juniper (Sabina przewalskii) are shown in the Vegetation Atlas of China 23 (Scale 1: 1,000,000) ca 15 to 30 km north of Fengtai . Until recently palaeoenvironmental data from the region were very scarce. The ear lier records obviously lack radiocarbon dates and their use for the discussion of the short term climate and environmental changes of the Bronze Age period is lirnited. Mid Holocene palaeoenvironmental records summarised by Chinese authors suggest that the period between ca 2550 and 2050 BC was wetter and warmer than at present and the 24 number of archaeological sites in North China increased notably at that time . Between 2635-2295 BC pine pollen dominated the pollen spectra from Haiyuan Ruin (36°15 1 N, 105°40 1 E) located in the present-day semi-arid steppe, and Qinghai Lake (36°54 1 N, 100°11 1 E) experienced high water levels. However, around 1000-1100 BC the level of 25 Qinghai Lake dropped significantly, suggesting a sharp aridisation of the climate . A recently published palaeoclimatic interpretation of the Holocene sedimentary, pollen and oxygen isotopic data from the Hongshui River section (38°10 146" N, 102°45 153" E, 1,460 m a. s. 1.), Tengger Desert, suggests several notable changes in tem 26 perature and precipitation in the region • Three phases of temperature decrease, coincid ing with the advances of glaciers in the Qilian Mountains and on the Tibetan Plateau, have been reconstructed at 3120-2720, 2350- 1790, 1560-1280 cal BC. The latter phase with relatively cool and moist conditions in the Tengger Desert was followed by a phase 27 of extreme aridity centred around 1050 cal BC . Another Holocene pollen record from the recently dried-up Eastem Juyan lake (41.89° N, 101.85° E, 892 m) in the Alashan Gobi suggests that a climatic phase wetter than at present occurred there between ca 1250 and 850 cal BC interrupting the time in 28 terval with relatively dry conditions . This wetter phase coincides with the radiocarbon dated time interval during which the excavated settlement existed at Fengtai. 23 Vegetation Atlas of China 2001. 24 Shi et al. 1993. 25 Shi et al. 1993. 26 Zhang et al . 2000. 27 Zhang et al . 2000. 28 Herzschuh et al. 2003. Sw-vey at Fengtai 2001 105 Radiocarbon age, Calibrated age, yr BC Calibrated age, yr BC Laboratory No. Material Location uncal. yr BP (probability 68.2%) (probability 95.4%) Dulan Nuomuhong 96°24'E, ZK 0061 wooden post 3670±90 2200-2160, 2150-1920 2350-1750 36°27'N Xunh ua Ahalta Hili Tuolongdu- 1810-1680, 1670-1660, BK 81027 M207 wood ash 3450±80 1950-1520 men 102°24'E, 35 ' 48'N 1650-1 640 Xunhua Ahalta Hili Tuolongdu- BK 81026 M12 wood ash 3230±80 1610-14 10 1690-1370, 1360-1310 men 102°24 'E, 35°48'N 1370-1340, 1320- Guinan Guantang 100°44'E, ZK 1326 coffin wood 2980±75 111 0,11 00-1080, 1060- 1410-1000 35°35'N 1050 Xunh ua Ahalta Hili Tuolongdu- BK 81030 M256 wood ash 2860±100 1210-900 1400-800 men 102°24'E, 35°48'N Guinan Jiatuhu 100°44'E, BK 77055 wood ash 2930±90 1270-1000 1390-900 35°35'N Datong Shangsunjiazhai BK 80014 M1046 coffin wood 2930±80 1260-1230, 1220-1000 1380-1330 , 1320-910 101°40'E, 36°56'N Datong Shangsunjiazhai 121 0- 1200, 11 90-1170, BK 77013 coffin wood 2860±90 1300-820 101 °40'E, 36°56'N 1160-1140, 11 30-900 Dulan Nuomuhong 96°24'E, ZK 0062 woolen cloth 2720±115 1050-780 1300-500 36°27'N Guinan Dayukou 100°44'E, ZK 1325 coffin wood 2720±75 970-960, 930-800 1070-760 35°35'N Huangzhong Panjialiang 980-750, 690-660, 630- ZK 1327 M78 human bones 2660±105 1050-400 101• 3o'E, 35•3o'N 590, 580-560 Xunhua Ahalta Hili Tu olongdu- 920-750, 690-660, 650- BK 81028 M158 wood ash 2630±100 1000-400 men 102°24'E, 35°48'N 540 Guide Shanpingtai 101 °27'E, ZK 1105 M19 wood 2640±70 900-760 980-750, 720-520 36°2'N Guide Shanpingtai 101 °27'E, 890-880, 840-750, 690- ZK 1107 M44 wood 2610±60 910-520 36°2'N 660, 630-590, 580-560 Datong Shangsunjiazhai BK 800 13 M1042 coffin wood 2580±80 830-750, 700-540 900-870, 860-410 101 °40'E, 36°56'N Datong Shangsunjiazhai BK 80011 M979 coffin wood 2510±80 800-520 800-400 101 °40'E, 36°56'N Datong Shangsunjiazhai 760-680, 670-640, 590- BK 80012 M989 coffin wood 2430±80 790-390 101°40'E, 36°56'N 580, 550-400 Huangyuan Dahuacun 101 °18'E, ZK 1323 M coffin wood 1780±80 130-350 AD 70-430 AD 36°42'N Fig. 13. List of radiocarbon dates for the Kayue Culture (adapted from: Radiocarbon Dates 1991, 285 - 289). Discussion and conclusions First results of the excavation at Fengtai suggest that the excavated part of the Bronze Age settlement existed for approximately 400 years, the last 200 years of second mil BC and 200 years of the first mil BC (e. g. 1000 ± 200 cal BC). The radiocarbon dates we have obtained (Fig. 8) help to narrow down the time interval in which the Kayue Culture is usually placed (Fig. 13 ). Both sedimentary and chronology records correspond with each other, helping to divide the history of the opened settlement into two periods. This conclusion is supported by the pottery complex. The lower layers of trench TG 1 attrib uted to the Early Pe1iod (1190-920 cal BC) only provide vessel and decoration types characteristic for the Kayue Culture. The upper layers of trench TG 1 and the cultural deposits of trench TG 2, attributed to the Late Period (980-750 cal BC), in addition to Kayue types also reveal Xindian types of the Zhangjiazui Phase. However, our dating for this period only partly support the dating suggested by Shui T. for the Zhangjiazui Phase. The results of the first excavation at Fengtai supp01t the hypothesis that the Xindian Cul ture gradually expanded from the banks of the Yellow River northward into the area oc cupied by the people of the Kayue Culture. At Fengtai we found three modes of house construction, which can be described as follows. The lower cultural layers in trench TG 1 are characterised by a steep slope, a 106 Xu Xinguo u. a. tremendous amount of ash, burned wood and pieces of mud-plaster but the absence of dwelling floors. This indicates a distinct mode of house construction built against the slope which is laid out on two ( or more) levels with timber as the main construction mate rial. Tue upper level rnight have been organised with a veranda open to the east, i. e. the Beichuan River Valley, constructed and supported by a woodframe system utilizing col umns and beams. Tue rear side of the house built into the slope rests on base sediment as shown in approximately 3 m depth in columns A-B (Fig. 3). Tue lower level can be imagined as having accommodated the slope leaving only an open space under the hori zontal wooden floor of the upper room and/or the veranda. Tue room or gap rnight have been used for storage and livestock and has gradually been filled with discard and refuse as indicated by the layers in column G at 6.0 to 7.0 m depth (Fig. 3). Whether the lower level was closed by walls could not be deterrnined although the barrier in columns G- H hintat it. Present day analogues to such a kind of house construction can be found in si rnilar landscapes and climate conditions, i. e. mountainous environments, for example in Nuristan, a very remote area of NE Afghanistan. The adequate vernacular response to the 29 steepness of terrain , Nuristani multi-storey houses exhibit features like those derived from our excavation results at Fengtai. The absolute age of the timber-house period at Fengtai, i. e. the age of the lower cultural layers, is about 1190 to 920 cal BC. After the devastation of the timber houses by fire and a levelling of the ground the site witnessed a second construction period but with a different type of housing. The pro files show horizontal house floors with multi-layered mud- and lime-plasters (Fig. 3, col umns E.F; 4, columns C - E about 2.7 to 2.9 m deep). By extending the trench at this level beyond the lirnits of the house floor to the surrounding footstep-hardened dwelling surfa ce the grove-like wall-base could be detected. However, no post holes or any other traces of wall construction material were found. Tue third architectural feature of the settlement found in the uppermost of the ancient cultural layers at a depth of about 1 m (Fig. 3, columns A.B) is the collapsed mud brick wall. Tue suggested age of the wall is 930-820 cal BC. We found that the wall was not a fortifi cation enclosing the entire settlement, but more likely the wall of a single building or yard. Surnmarising the results, we attribute the Early Period of wood house construction to the Kayue Culture. Tue construction of mud-floor houses and a mud brick wall is syn chronous to the spread of the Xindian Culture. Discussion of the human and environmental history of Fengtai valley must take the present-day conditions into account. Our analysis of the modern climatic variables sug gest that the moisture deficit is the main lirniting factor for tree growing in Fengtai area. However, even the present conditions support the growth of planted deciduous and even coniferous tree species in certain locations. Tue assumption that the region at the period between 1500 and 900 cal BC experienced slightly wetter conditions than at present is in agreement with the archaeological results. Tue features of house construction of the Early Period suggest that the area around Fengtai was at least partly forested during that time. This hypothesis of a wetter climate coincides with the timber-house period. However, it needs to be carefully checked by pollen and plant macrofossil analyses. Palaeoenviron mental records from NW China suggest dramatic changes towards aridity, which occur red in the region at about 1000 cal BC. Prelirninary excavation results from Fengtai dis cussed in the present paper propose a possible break in the lifestyle of the settlers (e. g. 29 Oliver 1998, 136f. 1007; 1017. Survey at Fengtai 200 1 107 abundant burned wood fragments, change in the house construction style) at around this time. Thus, additional data from the Fengtai area is needed to correlate the complex ar chaeological records with the palaeoenvironmental data from NW China. The results of our first excavation campaign in 2001 contribute substantial data to the discussion of 30 cultural contacts between Qinghai and Xinjiang dming the late Bronze Age . References Andersson 1925 J. G. Andersson, Preliminary Report on Archaeological Research in Kansu. The Geological Survey of China, Series A number 5 (Peking 1925). Bronk Ramsey 2000 C. Bronk Ramsey, Radiocarbon calibration and analysis of stratigraphy. Radiocarbon 37, 2000, 425-430. Debain-Francfort 2001 C. Debain-Francfort, Xinjiang and Northwestern China around 1000 BC. Cultural Contacts and Trans mi ss ions. In: R. Eichmann/H. Parzinger (ed.): Migration und Kulturtransfer. Der Wandel vorder- und zentralasiatischer Kulturen im Umbruch vom 2. zum 1. vorchristlichen Jahrtausend. Akten des Inter nationalen Kolloquiums Berlin, 23. bis 26. November 1999. Kolloquien zur Vor- und Frühgeschi chte Bd. 6, (B onn 2001 ) 57-70. Derbyshire et al. 199 1 E. Derbyshire/J. Wang/Z. Jin/A. Billard/Y. Egels/M. Kasser/D. K. C. Jones/T. Muxart/L. Owen, Land slides in the Gansu Loess of China. In: S. Okuda/A. Rapp/Z. Linyuan (eds.), Loess-Geomorphological Hazards and Processes. Catena Supplement 20 (Crernlingen-Destedt 1991) 119 - 145. Herzschuh et al. 2003 U. Herzschuh/P. Tarasov/K. Hartmann/B. Wünnemann, Early to mid-Holocene clirnatic changes in the Alashan Gobi, NW China: Conclusions from fossil and modern pollen spectra. In: S. Mischke/B. Wün• nemann/F. Riede! (eds.), International Symposium Environmental Change in Central Asia March 10- 15, Freie Universität Berlin Abstracts. Berliner Paläobiologische Abhandlungen 2 (2003) 45 f. Krieg et al. 1998 R. Krieg/J.-K. Liu/G. Müller/M. Schädler/E. Stemfeldt, Provinzporträts der VR China. Geographie, Wirtschaft, Gesellschaft. Mitteilungen des Instituts für Asienkunde 289 (Hamburg 1998). Leemans/Crarner 199 1 R. Leemans/W. Cramer, The IIASA CLimate Database for Mean Monthly Values of Temperature, Pre cipitation and Cloudiness on a Global Terrestrial Grid. Laxenburg Intern ational Institute of Applied Sys tem Analysis, RR-91-18 (Laxenburg 1991 ). Mei 2000 J. J. Mei, Copper and bronze metallurgy in late prehistoric Xinjiang. Its cultural context and relationship with neighbow·ing regions. British Archaeological Reports International Series 865 (Oxford 2000). Oliver 1998 P. Oliver (ed.), Encyclopedia of vemacular architecture of the world (Cambridge 1998). Pak 1996 Y. Pak, A study of the Bronze Age Cultures in the N01them zone of China. University Microfilms Inter national Dissertation Services (Ann Arbor 1996). Prentice et al. 1992 I. C. Prentice/W. Cramer/S. P. Harrison/R. Leemans/R. A. Monserud/A. M. Solomon, A global biome model based on plant physiology and dominance, soil properti es and climate. Journal of Biogeography 19, 1992, 117 - 134. Radiocarbon Dates 1991 Zhongguo shehui kexueyuan kaogu yanjiusuo (ed.), Zhongguo kaoguxue zhong tan shisi niandai shujuji 1965- 1991 [Radiocarbon dates in Chinese archaeology 1965- 1991] (Beijing 1991). 30 Mei 2000, 65. 108 Xu Xinguo u. a. Shi et al. 1993 Y. Shi/Z. Kong/S. Wang/L. Tang/F. Wang/f. Yao/X. Zhao/P. Zhang/S. Shi, Mid-Holocene climates and environments in China. Global and Planetary Change 7, 1993, 219 - 233. Shui 2001 T. Shui, Zhongguo xibei diqu qingtong shidai kaogu lunji [Papers on the Bronze Age Archaeology of Northwest China] (Beijing 2001). Stuiver et al. 1998 M. Stuiver/P. J. Rei.mer/E. Bard/J. W. Beck/G. S. Burr/K. A. Hughen/ B. Kromer/F. G. McCormac/ J. v. d. Plicht/M. Spurk, INTCAL98 Radiocarbon Age Calibration, 24,000-0 cal. BP. Radiocarbon 40, 1998, 1041-1083. Vegetation Atlas of China 2001 X. Hou (ed.), Vegetation Atlas of China (Scale 1: 1,000,000). Science Press (Beijing 2001). Wagner 2001 M. Wagner, Kayue - ein Fundkomplex des 2. Jahrtausends v. Chr. am Nordwestrand des chinesischen Zentralreiches. In: R. Eichmann/H. Parzinger (ed.): Migration und Kulturtransfer. Der Wandel vorder und zentralasiatischer Kulturen im Umbruch vom 2. zum 1. vorchristlichen Jahrtausend. Akten des In ternationalen Kolloquiums Berlin, 23. bis 26. November 1999. Kolloquien zur Vor- und Frühgeschichte Bd. 6 (Bonn 2001) 37-56. Xie 1980 D. J. Xie, Gansu Yongjing Zhangjiazui yu Jijiachuan yizhi de fajue [Excavations of the Zhangjiazui and Jijiachuan sites at Yongjing County in Gansu Province]. Kaogu xuebao 2, 1980, 187 - 220. Xu 1988 X. G. Xu, Shilun Kayue wenhua de leixing yu fenqi (1) [Towards sub-groups and periods of the Kayue Culture]. Qinghai wenwu 1, 1988, 35-44. Xu 1989 X. G. Xu, Shilun Kayue wenhua de leixing yu fenqi (II) [Towards sub-groups and periods of the Kayue Culture]. Qinghai wenwu 2, 1989, 38-44. Xu 2002 X. G. Xu, Qinghai kaogu de huigu yu zhanwang [Qinghai archaeology in retrospect and prospect]. Kaogu 12, 2002, 3 - 11. Zhang et al. 2000 H. C. Zhang/Y. Z. Ma/B. Wünnemann/H.-J. Pachur, A Holocene climatic record from arid northwestem China. Palaeogeography, Palaeoclimatology, Palaeoecology 162, 2000, 389-401. Zhongguo wenwu dituji, Qinghai fence 1996 Guojia wenwuju (ed.), Zhongguo wenwu dituji, Qinghai fence [Atlas of Chinese cultural relics, vol. Qinghai] (Beijing 1996). Mayke Wagner Wang Wei Karl-Uwe Heussner Niu Shishan Bettina Griess Song Jiangning Deutsches Archaeologisches Institut Archaeological Institute of the Eurasien-Abteilung Chinese Academy of Social Sciences Im Dol 2-6 27 Wangfujing Road D-14195 Berlin 100710 Beijing P. R. China Xu Xinguo Pavel E. Tarasov Xiao Yongming Alfred-Wegener-Institut for Cai Linha i Polar and Marine Research Archaeological Institute of Qinghai Province Telegrafenberg A43 43 Weimin Road D-14473 Potsdam 810007 Xining [email protected] P.R.China S urvey at Fengtai 2001 109 wl'HJ:~$i:i'~:W:2001if:~-t~ffil~lf:1Jt~~*i5- iti=ffioo ®r, .rft :lt~, MaykeWagner,Karl-UweHeussner,PavelE.Tarasov, Bettina Griess, Berlin, 41:t!: w 5Rllr :lt~, ~ kaA ~*~ ®r 51§ Wl'HJ:~'fü:-=tW~ra'J@:* :ltm ~;f □ ©(JE *~@:*mm~ , ;1~:riJFJl'.A~#ii9J%□ 4: f~ ~ :lfxt:Ef :IJt~ n1~ B<:l *~tm IR o ~-t ~r.m• l3jJ , 0 JG mr 4oooif ir:;{5 a Jt *s B{J~Jn:rEt;a;: ~~ili~78~~ffi~ft~~--o~m~EX~~fil13jJB~~r~@:ffi1FSm~ tllilR, ffjj~Jtfi/J!l~5t;;tai:i-)i!;t, [[!Jli, ffiijj , ®~J'J~®f1:PfU1rmffß, llfl-W"~AA ffl~~M.~IR~EX~~~~.A~ffl~tto ~§f!TÄ~, ~r~E~tt~~~~ fflffl*~Jl'.~~w,m~m~Ex~~~•~••, ~-:lf~•~•~&~~~M ~B%AiR o ~~x~~-l'HJ:~>Fffß~IR~- ~fffflX~o~#ff*fflA~&~~~ffiMMm Jt~T~~~~~, *Wffl~ft%~~&-~ßif·ffl-~7j;::~if~, @~f!T~ &~mff~- ~~~x~~--~~*fi*~~~~ffi~ffio ~oo~m~~Elfmß~fi•13jj, *M~~mM~4:~~~mm~~ffittt~*~ --~tl~~7~~om$~~hl~~E~ ~ii"lf:IJt~Jl'.-~~rW~~~~~-t tllilR••;f □ tm~fl~tlB{Jfü~ o ~:!Jl:,ffi?,l ~~fffflB1·f~~--fü:r1iJl)J~$~tf>F, :ltße® r460mo 1980:tJ':{~ , ffl'HJ:~' ~EPfi~f.fü 7~~~f>JX1-t~t1L 2oooif Wl'HJ:~'~EPfi;f □~ OOt±~;f4~ fBt~Ef!Jf71'. Mffl~,MOO~E~®°(lr~EM~~m~~hlmfi7Wm~~,W01 :tJ':6~*fi7~ ~ÄM- -t-~~~-, ~~§~~-~~-- ~~~ ~~ ' TMm~•mtt#ffitt *4ir-A o ~ffil~:'= J!l!Jl!ffUR ~ ~ tftll TWi-t-~ )ZJ o ~1ZJ1 (;Fiz:l:itftiiif) ~~1ZJÄlEm:lt1J(,:;J, ~2, -lf8* o tt*ttLl iifB{J...l:ffßl.3*~ J:-t-1:t!:fö:&~föi;"IJ Mr~1ZJ~; Jtroo~~~~±•llirn~~±~•. ~1r~~*•l3jj~-t-±~•~ tlfä 1J,, PJ~~~-=t--t- 1J,ru!.ffl, JJ,l(JJiti~WWif.fflÄ 930 -820 cal BC, ±:1:f:i:lMr oo~~-t-~±m, Jt~~~*~~~•(,:i]r~Md&, ~~-t-~±mPJ~~~±~ •~~M~E~iif;a~-t-E~ iif~r~A-t-•~~n~~~~$~~•mmm~ ~~~*•B%~±!1lJH_jL ~Jt, i9J40~fm;f□~~ . a 2.2-2.3 *% □ 3.1 *~iiaJJ~ ~ey~*~*~*~~-ß~~±m,~®$fflPJ~~*7•-~~~~MMttffio a*®~oo~2~-2.9*~~-~~Tm~$~~mmmttoo,Jt ~*~*•~~ ~~~, PJ~•l3jj~m~m~~m~~•~oo~•~· a~-t-~ffl~~~Ttthl~ ;j11Jjij, M:J:±t.tl:~71tl:B{J*m ~nJl:M·t1:UWif•ffl~}-,A 980-820 cal BCi lJ11 30-930 cal Be; riif~ - -t-~~*ii*ml□ ~±~ffttmm$t9', a1Ftt 1J iiffü:r3 .8 I IJ4.3*r*Ji, rm:t±®tt1Jiif JnlJ1.flr3.4-3.7*r*ii, - :l}c*1-t~ttlli-~nJi:Mtl:~.OO Ll ifii&ffl Ä 1320- 1110 cal BC, ~ rB{Jmfü:~~~$~B{J*~, X1-tma~~:it:7.12*o ~1ZJ2 cmi ooo) ~-t-~1ZJ~ft ITT ffl~7~~-mM%~~ffiiiio•-mÄ~~&•~~ß~± •~•±; • =mÄM~±; • =m*~~±,~±,*~· Jt ~~~~~ffii9J~ 110 Xu Xinguo u. a. itfm, :litBIAä EI Er-J~'G1Hi%Pts11l'®.f{/l'r, ~~~J.ml!J•, ftfüiAÄJ!m¾ - ~ i-&:xx~W81/% fü , W31-*~ f:Fiß 81JJJIJJt·t1~~:q::~ *¾ 830-765 cal BC o ~~W31-~itJ81i1U9'~fil, "i:iJ ~~%-~-3/.i?f ÄW31-~fl-~: ~11;J1813-7* 81Xft~:q::1~Ä 1190- 920 cal BC, ~r l\!Wt ~ii;Jl 813*~_.t_, ~1tJ281:q:: 1~Ä 980 - 750 cal BC, ~rfljJt/'Jo 3~, x1~1-~±ll::81%~:5tM~$~4' J§ ~7':1>\lfl 81 ~ 1lffi *%nlZ o ff#JJ-1§ ~~81~~~~#Ä~al!!Äft~~l81*~~~~81~#ft$,M~$ ~J1ijj # t'ß 1f'.i?t~ 81 mJ2t o ~~tt .zi@: 81 fl: %fä j> , ~ ~ ¾X>CEI= 7':. tD f;\jf(l:J ~ ; ~1ijj ~ - ~~~: ~~ffl@~Ä~, ~d>•81~m~: ~ßM~fflß~ffhß~ #~fflß, ~~ß~#Mft-~fäd>: a~fl:81Q$,~W81**~#~00 ~&re~_t•~•~ttß81~ffi: a•n_t~~ffim~tt81IB~~tt. ~lf ~tt,~tt,~m~81ti~tt~tt~omffl~~-81~~fil·~-$~~fil "i:iJ~~~B?tÄl\!~W3M, l\!M$Äffi~~, Q$ffl-~*•*· -~#~ 81ffire~ma~~m~~. ~fim~~tt; ~M~~re~m~anw81•* ~1iN., ~tJJ □ ~tti-&:1 □ 1~ ~0 ~;IJtiJ~E~~ * x1r*~~~~~,~~~ffi~~~M~81~M~~mriA~J!m81~~ *X1-t~xL $ ~1i;J ¾$!-~ 81i-&:lriJ7.)(f1'PXJJ~ Bi Ll1 l'äJ1~~, 33:mB'-J±-tMfl~ ,8~7T T ±m~~81M~o rnr•±~~-ffiaffl=~h~±_l_ffi, 3~*•A~ J§h;J<,5±81*®~1~1~m, ~J'h~IJE9jt±81Nf$, rn.tl:t~l~1~1itJ1t, ~B1, ~ -"i:iJ~~¾~~~1i~81~-1-l!l*oM~~~~~fil~~M~~ ~M~ 81M1i~OOr81mJ2t, 33:#~~~~81~mxWJ:iJ~¾~~•3/.i~a81- -t ~~@:l!I 0 x1~ffrttt~~1~:PJfn81~**SJj, ft 2550 - 2050 Bc:q::räJ~ {~ U~fJJlft1MiB11~ WJ, 1.§.¾ 1100- 1000 BC w~1i'iJHi~fü T7.KN~~r~81Mlj[ o ftia:~*81ffi $~~fil•~n~**~· amm~•a~~ffittt~füT~~~~81~&~ ~J.)(~1-t, Ä~fr 3120 - 2720, 2350-1790, 1560-1280BC1f - ~~1ßr\,x-j" fH~i!w 811JJ(}ll1fj(, ft 1050 BC 1W J§~ - l.X5la?.!~ 81-=f ~JJ.14 o ~ - 1-~ifr ttt81ffi$~**~, 1250 - 850 BC ~ - ,t-;t§x-j"y~~iJ81~fl-~, IEx1J.11Zr$~ ~±ll::~1!tElM1t~s11lfr~o i-fit.!§;~it $ ~ ~±11:: 81 ffl - ljz~flffi* SJl 33: 1-wffl 811~ 81 .3/.i fil~ T 4OO:q::, ±i!!~*lD ~ ~*1c:5RW*~ B~~tftEl81$:5t "i:iJ~ :lW?tÄl\!~W31-~fl-~, l\!M C1190 - 920 cal BC) ~r-tt"JX1-t, ßfM (980-750 cal BC)~T~-tt"JX{-t81~fil:>'r, ~~7f,T*mx~**•~~81~filo~m•~~~*mx~M-~~~~ ::lttU3zs:-tt"JX 1-t?t11J ~ 811fi ~ o ft~ii;Jl815.5 - 5.7*~~. *~ft$'i'81~{.fCf$~± T- :tk:•81 Q$~ #, %~ffl~81~fl:ft7($~a,ffi*~'*~~n~••rwlf~~o l!I :;0J!#B%~-t~x~81n~**x~81~~~•~•~m81•*· ~AiA ÄJK¾-t~x~l\!M81~~~~oM~#~N_t~, J!~~Ba$~mhl81 ~mN•~~TJ!#~~o~M ~fl:81-~,%~W&~T*mx~81~fil, amhl_t~~~*mm:~~81-ffl~~%~81~~~~~•*~T*mx~ 13"-J~0~, 'f::{i'H§ 3T*[i§j({-tat:Jt,)jllfrm81~~* nt~~. :q::1~Ä 1100 - 800 BC o ft$~mhl~~81/%~•mmJ2t"i:iJ~?tÄ = #offl- #¾*~•m, frffi 3r-t~x~B1M81rWx~m~~~81*• 81~~,'.i?tM81*~$ • ~~. SLrrVey at Fengtai 2001 111 ~~~~~~~~~*~mm, @fü*~r*~; ~~~mm~~~~~~ ~~T$m• ~~~~~m~mttoo~a•oo~&~~~M, @~~~~tt ~, ~~ffi=#mm: ~•~oo~x~m~~Tffl~~±~~, ~~ffi - # mm, füffiWHrr mm*:r-tt§ 3 T $frsX1-t ~ ~U't Ji: B,J ;J!] 0 0~ $&ffl hl~~~A~*~ ~ fü~~~~~Ä~~~~ o ~~~~~~ #~*~~N~ft~~ ~W*~*~ ~~ ~~,@~~~~~ ~ ~N~~,~ .mm~t'iJ~~*~ritImJrJt#tottrit# o~~ {~:UJfYi~~*1soo-9ooto 1190- 920 cal BC~~ ~fü T ~~ $x11fil ~I] Btl tk: ~ Z--f if ~ ~ ~ :f,!ll, ~ :n 1t! JA ~ir ~ Btl *tiJ Yl ffl ~~.3 ~~~pJ~~-#~~o_3~ , ~@ ~~~fi~~ill- ~~ffi~~ ÄB~~~:UJfYi*i.iEi:!J3 o (Translated by Jin Gui-Yun) Pe3l0Me ITepBble p e3yJJbTaTbl paCKOOOK 2001 ro.D,a B